NetNews Usenet Archive 1992 #31/NN_1992_31.iso/spool/comp/sys/ibm/pc/digest/186

Internet Message Format | 1992-12-22 | 23.9 KB

Path: sparky!uunet!spool.mu.edu!agate!ucbvax!WSMR-SIMTEL20.ARMY.MIL!Info-IBMPC
From: [e-mail] ("Info-IBMPC Digest")
Newsgroups: comp.sys.ibm.pc.digest
Subject: Info-IBMPC Digest V92 #196
Message-ID: <[e-mail]>
Date: 22 Dec 92 07:03:31 GMT
Sender: [e-mail]
Reply-To: [e-mail]
Distribution: world
Organization: The Internet
Lines: 635
Approved: [e-mail]
X-Unparsable-Date: Sat, 19 Dec 92 16:44:13 GMT+1

Info-IBMPC Digest           Sat, 19 Dec 92       Volume 92 : Issue 196

Today's Editor:
         Gregory Hicks - Rota Spain <[e-mail]>

Today's Topics:
                     486dx, 486sx, 486dx2 (5 msgs)
                      [TDR] 'Burping' Hard Drives
                     [TDR] Re:  Redundant CPU Chips
                 A problem with DOS 5's EDIT under WIN3
                           auditing software?
                      Cyrix 486 CPU'S. Some facts
                          Frozen Hard Disk (?)
                           Speach Recognition
                  DMA Transfers with MC-IEEE-488 Board
                        mktime() adjustment bug
                   Network printer from PCs and MACs
                     PD1:<MSDOS.DATABASE>/GC11.ZIP
        PS/2 High Density mode for 3.6" Floppy isn't Compatible
                 Winmarks, Specmarcs, and all that jazz

Send Replies or notes for publication to: <[e-mail]>

Send requests of an administrative nature (addition to, deletion from
the distribution list, et al) to: <[e-mail]>

Addition and Deletion requests for UK readers should be sent to:
<[e-mail]>

Archives of past issues of the Info-IBMPC Digest are available by FTP
ONLY from WSMR-SIMTEL20.ARMY.MIL in directory PD2:<ARCHIVES.IBMPC>.

----------------------------------------------------------------------

Date: Wed, 9 Dec 92 16:12:03 GMT+7
From: "Douglas R. Nebeker" <$[e-mail]>
Subject: 486dx, 486sx, 486dx2

> >   What is the difference between the 486DX, 486DX2,
> >and 486SX with the Overdrive chip installed?

[ text deleted ]

> A 486DX is full 486 computer, yes it has a math coprocessor built in
to it but
> it is a limited math coprocessor. There is a socket on all 486DX
mother boards
> that I have seen for a math coprocessor or over drive chip. I am not
totally
> sure but I believe that the over drive chip is the same thing as a
math
> coprocessor.  The difference between a 486DX and a 486DX2 is that a
486DX
 50MHZ
> runs the whole system at 50 MHZ and a 486DX2-50 would run the system
at 25 MHZ
> and the speed of the processor is doubled so the processor runs at
50MHZ.
>
> A 486SX is not a full 32 bit computer.  The memory I believe is only
16 bits
> when the processors is 32 bits.

The 486SX IS a FULL 32 BIT computer (this SX meaning 'half computer'
stuff applies ONLY to the 386).

A 486DX is a full blown 486.  A 486SX is a full 486 without the math
coprocessor.  A 486DX2 is a full blown 486 with the system board
running at half the advertised rate while the CPU runs AT the
advertised rate (if you have a 486SX-25, you can put a 486DX2-50 in and
the rest of your system will still run at 25 MHz while your CPU is
running at 50 MHz).

and as I understand it, the OverDrive chip is simply a DX2 chip sold by
Intel (Intel sells OverDrive, other vendors sell DX2, but they're the
same chip)

Douglas R. Nebeker                     Internet: $[e-mail]
Brigham Young University

------------------------------

Date: Wed, 9 Dec 92 17:25:00 EST
From: [e-mail]
Subject: 486dx, 486sx, 486dx2

X-News: urvax bit.listserv.pctech-l:5327

>From: [e-mail]
>Date: Thu, 3 Dec 1992 15:33:00 CST

>   What is the difference between the 486DX, 486DX2, and 486SX with
>the Overdrive chip installed?  I know the 486SX with the Overdrive chip
>makes use of the math coprocessor socket, so does installing one mean
>you no longer need a math coprocessor or just that you simply can't
>have one?

>   Also, can an Overdrive chip be used on a 486DX computer?  I am
>under the impression that a 486DX computer has a built in math
>coprocessor, so there would not be a socket for one, no?

>   The 486DX2 is similar to a 486SX with the Overdrive chip installed,
>isn't it?  Or, does the 486DX2 still have a vacant socket for a math
>coprocessor?

A 486DX is full 486 computer, yes it has a math coprocessor built in to
it but it is a limited math coprocessor. There is a socket on all 486DX
mother boards that I have seen for a math coprocessor or over drive
chip. I am not totally sure but I believe that the over drive chip is
the same thing as a math coprocessor.  The difference between a 486DX
and a 486DX2 is that a 486DX 50MHZ runs the whole system at 50 MHZ and
a 486DX2-50 would run the system at 25 MHZ and the speed of the
processor is doubled so the processor runs at 50MHZ.

A 486SX is not a full 32 bit computer.  The memory I believe is only 16
bits when the processors is 32 bits.

exit

------------------------------

Date: Thu, 10 Dec 92 16:59:48 EDT
From: Paulo Francisco Sedrez <[e-mail]>
Subject: 486dx, 486sx, 486dx2

>Date: Wed, 9 Dec 92 17:25:00 EST
>From: [e-mail]
>
>>From: [e-mail]

>> [...Text deleted...]

>A 486DX is full 486 computer, yes it has a math coprocessor built in
>to it but it is a limited math coprocessor. There is a socket on all
>486DX mother boards

No, it is a full math coprocessor. The main advance from 386 to 486 was
the insertion of the math coprocessor into the main processor AND the
optimization of the instructions.

>that I have seen for a math coprocessor or over drive chip. I am not
>totally

Yes, you can use the Weitek coprocessor; it is faster then Intel's.

>sure but I believe that the over drive chip is the same thing as a
>math coprocessor.  The difference between a 486DX and a 486DX2 is that
>a 486DX 50MHZ

Not so. As I heard, over drive is the full 486DX2.

>runs the whole system at 50 MHZ and a 486DX2-50 would run the system
>at 25 MHZ and the speed of the processor is doubled so the processor
>runs at 50MHZ.

Nearly right. The double clocking technology is based on instructions
coexecution, i.e., it allows two instructions to execute at the same
clock cicle. The effect is a processor running at a speed similar to
the processor running at the double clock.

>A 486SX is not a full 32 bit computer.  The memory I believe is only 16 bits
>when the processors is 32 bits.

Wrong. The 486SX if a full 32 bit computer, but the math coprocessor is
not enabled. I don't now how Intel do this.

------------------------------

Date: Fri, 11 Dec 92 11:59:30 GMT+7
From: "Douglas R. Nebeker" <$[e-mail]>
Subject: 486sx, 486dx, 486dx2, etc.

> From:  David Andrew Vaughan
> >From: "Douglas R. Nebeker" <$[e-mail]>

> >A 486DX2 is a full blown 486 with the system board running at half
> >the advertised rate while the CPU runs AT the advertised rate (if
> >you have a 486SX-25, you can put a 486DX2-50 in and the rest of your
> >system will still run at 25 MHz while your CPU is running at 50 MHz).
>
>     Why is it advantageous to keep the rest of the system at 25 MHz?
>     If it is a simple matter of money, I understand.

    I believe it is just a cheaper upgrade path.

>
>     Is there a particular advantage to having a processor running at a
>     different speed than the rest of the system?  Is there some system
>     feature you can now exploit that was previously impossible when
>     everything ran at the same speed?

    No.  Basically a 486SX-25 is the cheapest 486 you can buy.  If
later you need more power, but can't afford a 486DX66, simply buy a
486DX2-50 chip and your overall system speed will increase (not quite
double--the CPU runs double, but all the system buses are still running
at 25MHz).

>     What does it mean to have "the rest of the system" running at
>     25 MHz?  Is that the bus, memory, the disk?

    "Rest of system"=all components that are clocked on the system,
i.e. the data bus, the address bus, the control lines, etc.

>     If the real horse power of the machine is the CPU, then what
>     difference does it make what speed the rest of the system is
>     running?

    The CPU actually does all the work, but it can only work on data as
fast as it gets it.  So, any IO to the CPU still takes place at the
older "slow" rate, while the CPU is screaming at double the "slow"
rate.  This accounts for the overall system being much quicker (perhaps
70% quicker) but not quite 100% faster.

Hope that's as clear as mud! ;)

Douglas R. Nebeker                     Internet: $[e-mail]
Brigham Young University

------------------------------

Date: Thu, 10 Dec 92 15:14:50 EDT
From: [e-mail]
Subject: Re: 486dx, 486sx, 486dx2

From:  David Andrew Vaughan
>Date:         Wed, 9 Dec 92 16:12:03 GMT+7
>From: "Douglas R. Nebeker" <$[e-mail]>

>A 486DX2 is a full blown 486 with the system board running at half
>the advertised rate while the CPU runs AT the advertised rate (if
>you have a 486SX-25, you can put a 486DX2-50 in and the rest of your
>system will still run at 25 MHz while your CPU is running at 50 MHz).

    Call me naive.  I'm no EE (as this message will prove).

    Why is it advantageous to keep the rest of the system at 25 MHz?
If it is a simple matter of money, I understand.

    Is there a particular advantage to having a processor running at a
different speed than the rest of the system?  Is there some system
feature you can now exploit that was previously impossible when
everything ran at the same speed?

    What does it mean to have "the rest of the system" running at 25
MHz?  Is that the bus, memory, the disk?

    If the real horse power of the machine is the CPU, then what
difference does it make what speed the rest of the system is running?

David Andrew Vaughan
GE Consulting Services

------------------------------

Date: Fri, 11 Dec 1992  21:25:27 EST
From: "Message Center" <[e-mail]>
Subject: 'Burping' Hard Drives

Tom Frenkel <[e-mail]> asks about his
'burping' hard drive."

Certain hard drive models have a feature called "Auto Park" in which
the drive, after a certain amount of inactivity, "parks" itself by
moving the heads to an unused area and aligns them so that they don't
get damaged.  This helps increase the lifespan of the hard drive and
reduces the amount of power it consumes (and, hopefully, the amount of
heat it produces).

The "burping sound" is the heads being retracted from the drive
platters when the drive decides to park itself.  The hard drive on my
computer (A Seagate 84meg SCSI drive whose code number (ST3104N?)
escapes me) does the same thing.  Three years ago when it did this, and
I had put myself into severe debt to purchase it, it scared the
excrement out of me for fear I'd gotten a lemon that was acting funny.
It wasn't until later that it occurred to me what was happening.  When
you spent the equivalent of a week's pay (even if it is on your
plastic) you get very worried if something is wrong.

Unless your computer tries to access the hard drive at the exact moment
it decides to park itself, this action should not cause your computer
to change its activity at all, i.e. it should not slow down your
computer.  If it is doing so, I'm at a loss to figure why.

--- Paul Robinson -- [e-mail] These opinions are mine alone

------------------------------

Date: Thu, 10 Dec 92 21:47:51 EST
From: "Paul Robinson" <[e-mail]>
Subject: Re:  Redundant CPU Chips

In reference to the Overdrive and 80486dx2 chips, the same situation
exists with respect to the 80486sx and the coprocessor chip, the
80487sx.

Intel has gotten around the problem of someone buying the "upgrade"
chip while still having the original chip present.  The upgrade chip
won't work without the original chip.

This is because certain circuitry in the upgrade chip relies on the
existence of the original chip.  Which upgrade circuitry?  The
circuitry that disables the original chip!

So what you have is a replacement CPU that its main purpose is to tell
the other CPU to turn itself off.  No, you can't use both CPUs or take
the original out.  Intel already thought of that.

The 80487sx is a full processor with a coprocessor built in which
disables the 80486sx.  I figure that if they did it once (with the sx)
they can do it again (with the dx).

Paul Robinson -- [e-mail]
These opinions are mine alone

------------------------------

Date: Sat, 12 Dec 92 08:53:40 CST
From: BF04000 <[e-mail]>
Subject: A problem with DOS 5's EDIT under WIN3

I use two machines: a 40mhz 386sx (Blue dolphin) and a 33mhz 486dx
(Flex) both using DOS 5.0 and Windows 3.1. The 486 uses 386MAX to
manage memory (8 meg) the other does not (4 meg).  Here is the problem:
I often use DOS' edit in a MS-DOS window, and every once ina while
(about 10% of the time) EDIT takes a long time to load - 10-20 seconds!
I call EDIT file.txt where file.txt is small (1000 bytes) and the
screen goes blank - the cursor sits in the lower left corner and
nothing happens - no disk activity - nothing for 10 or more seconds.
What could be causing this?  In config.sys I have files=50, buffers=30.

Chris K. Caldwell    [phone]     +--------------------------+
Math/Comp. Sci       [phone]     +       This space         +
Univ. of Tennessee at Martin            +           is             +
Martin, TN 38238 USA                    +        for rent          +
[e-mail]                     +--------------------------+

------------------------------

Date: 11 Dec 1992 10:06:38 -0700 (MST)
From: [e-mail]
Subject: auditing software?

Does anyone know of any cheap or shareware software for MS/DOS
compatible computers that can track program use?

We are running a (very) small computer lab at the university, here and
want to keep track of how often programs are being used and how often
computers are being used.  We are also interested in knowing how long
the programs are being used.

Thanks, in advance,
Nohl Lyons
Internet: [e-mail]
University of Arizona

------------------------------

Date: 10 Dec 92 00:58:52 GMT
From: [e-mail] (Eleftherio Simotas)
Subject: Cyrix 486 CPU'S. Some facts

I was recently interested in getting a motherboard, and came across the
Cyrix $86 CPU's.

I read the following stuff from BYTE.

The SLC is 16 bit external bus, the DLC is 32 bit.  Both have 1K cache.
Neither has math copro.

Cyrix bundles the chips with coprocessors for very little extra (~ $20)
(OEM' S).

The chips are supposed to implement the etxra 486 instructions.
(Presumably from what I read on the net this is not quite true.). I
believe though that anything that will not run on a Cyrix 486 will not
run on a 386 either.

The chips have a hardware multiplier (the i486 does not?) but lack
burst memory access. I think this fact can justify the performance
reports.

The Cyrix cpu's have a fast core (as fast a INTEL's 486?) and thus
perform faster than a 386. They lose out to the i486 since without
burst memory access, bandwidth to memory is far less (386 level).  On
floating point the difference should be more significant since the i486
has the coprocessor on chip and communicates with it over a bigger bus
with much smaller delays.

I think Cyrix did not make the best choice by giving these chips the
486 name. They were really meant to compete with high end 386
processors ( 33/40 MHz), not INTEL's 486.

If they are probably supported by the BIOS-motherboard combination ,
they present excellent value for money.  A good value for a 386/40 w
64k cache is around 180+85 = $265 (December's Computer Shopper).  In
the same issue I saw an ad for a Cx486DLC-33 MB with FPU for $300 (no
cache mentioned). If the MB included 64k external cache I think the
extra $35 would be more than worth it. (IMHO).

E. Simotas
EE, Rutgers University, NJ.
Rutgers University.

------------------------------

Date: Sun, 13 Dec 92 17:14:22 IST
From: Zvi Lando <[e-mail]>
Subject: Frozen Hard Disk (?)

I have a friend who has an old XT with a 20MB disk. He called me to
tell me his disk doesn't always boot. When I asked what he meant by
that, he explained. If the room is heated, the disk boots, if it is
cold, it won't. I've asked around and cannot find an answer to this,
though it probably means that his disk will soon pass on "to a better
world".

I told him to heat the room, boot the disk and copy all his data down,
which he did. Just for curiosity sake, I thought I'd ask the members if
they can explain it.
Thanks -
Zvi Lando


------------------------------

Date: Thu, 10 Dec 92 11:36:22 GMT
From: Arthur Laferriere <[e-mail]>
Subject: Speach Recognition

I'm interested in locating and evaluating state of the art
hardware/software for speech recognition with an IBMPC.  A couple years
ago PERX, a company in California had something that could be trained
to recognize some 100 or so spoken words.  They appear to have gone out
of business.  Any leads?

Thanks.  Art Laferriere, [e-mail]

------------------------------

Date: Fri, 11 DEC 92 17:27:24 BST
From: [e-mail]
Subject: DMA Transfers with MC-IEEE-488 Board

        I have got an MC-IEEE488 Scientific Solution board in my IBM
PS2/70A, and trying to transfer data through DMA for fast data
transfer. I don't have any software to enable me to do this. Does
anybody have such a software to make it possible? or could provide some
information that I could get in touch with Scientific Solution in case
they may have such a software.   The software which Scientific Solution
provide does not use DMA (it is written in Basic), so any guidance on
modifying their code to use DMA would be much appreciated.

                                        Parviz Fozooni.
                                        RHBNC, University of London.

------------------------------

Date: Thu, 10 Dec 92 12:29:18 EST
From: Jim Van Zandt <[e-mail]>
Subject: mktime() adjustment bug

According to Borland documentation, mktime() accepts a pointer to a
struct tm in which the fields are not restricted to their normal
ranges.  "If the fields are not in their proper ranges, they are
adjusted." Presumably, the intent is to allow arithmetic on dates
(e.g.: an 18 month loan signed today comes due on what date?).
However, the Borland implementation does not correctly adjust the
fields.

In the following, hour, minute, and second fields were set to zero.
The initial dates in each pair differ by ONE DAY.  The dates after
adjustment by mktime() differ by up to ELEVEN YEARS.  In the printout,
the month offset has been removed, so 1 = January and -1 = November of
previous year.

     before            after
--------------    --------------
year month day    year month day

1998  -26    0 -> 2017   10   23
1998  -26    1 -> 1996    2   29

1992  -50    0 -> 2009   10   23
1992  -50    1 -> 1988    2   29

1990   54  -58 -> 1994    4    2
1990   54  -57 -> 1994    4    4

1990   -1  -70 -> 2011    8   14        (neither of these is right -
1990   -1  -69 -> 2011   10   24        the real dates are in 1989)

1990  -33   48 -> 1988    2    7
1990  -33   49 -> 1996    3   21

1986   16 -156 -> 1986   10   26
1986   16 -155 -> 1986   10   26

1983  -45   48 -> 1980    2    7
1983  -45   49 -> 1988    3   21

1982   5   -8 -> 1982   4   21
1982   5   -7 -> 1982   4   22
1982   5   -6 -> 1982   4   23
1982   5   -5 -> 1982   4   24 <-- off by one
1982   5   -4 -> 1982   4   26 <-- correct
1982   5   -3 -> 1982   4   27
1982   5   -2 -> 1982   4   28
1982   5   -1 -> 1982   4   29
1982   5    0 -> 1982   4   30
1982   5    1 -> 1982   5    1
1982   5    2 -> 1982   5    2
1982   5    3 -> 1982   5    3
----------------------------------

I'm using Borland C++  Version 3.1, small memory model.

The mktime() with Borland C++ Version 3.00 makes all these errors plus
several more, including:

1985   22   26 -> 1986   10   26
1985   22   27 -> 1986   10   26
----------------------------------

               - Jim Van Zandt <[e-mail]>

------------------------------

Date: Fri, 11 Dec 1992 14:29 PST
From: "Todd W. Schmitzer" <[e-mail]>
Subject: Network printer from PCs and MACs

Hello,

        I am looking for software that will redirect printing on a PC
and a MAC to a LPD server.  I am hoping to find a LPR program for both
a PC and a MAC the will transparently redirect to the LPD server (ie
for the PC, redirects LPT1 to LPR).

        Does anyone know of software to do this?  I'd prefer to find
something in the public domain, but I'll look at commercial software
also.

        Does anyone know of other alternatives to LPR/LPR to provide
transparent network printing to the same printer from both a PC and a
MAC?

        Thanks in advance.

Todd Schmitzer
Library Systems/Services Manager,
Data Communications and Networking Specialist
Academic Computing Center
Santa Clara University

------------------------------

Date: Sat, 12 Dec 1992 00:49:43 -0500 (EST)
From: Keith Petersen - MACA WSMR <[e-mail]>
Subject: PD1:<MSDOS.DATABASE>/GC11.ZIP

[e-mail] (Neal Baldwin) wrote:
> I FTP'd the GC11.ZIP file, but got a crc error when I unzipped it. I
> was in binary mode at the time . A colleague of mine tried to FTP it from
> OAK.Oakland.Edu, and got the same error. Are we doing something fundamentally
> stupid, or is ther a problem with that zip file? (we're both using pkunzip)

Neal, what version of PKUNZIP are you using? Most recent ZIP files on
SIMTEL20 require PKUNZIP version 1.10, which has been available for two
years.  Info-ZIP's free UNZIP50.EXE is also available from SIMTEL20 and
OAK.Oakland.Edu.

Script started on Sat Dec 12 00:39:08 1992

$ cd /pub/msdos/database
$ unzip -tU gc11
  Testing: gc.exe        OK
  Testing: gc.txt        OK

$ unzip -vU gc11
 Length  Method   Size  Ratio   Date    Time   CRC-32     Name
 ------  ------   ----  -----   ----    ----   ------     ----
 145894  Implode  56250  61%  11-07-92  16:25  69b43c7d   gc.exe
  11150  Implode   4016  64%  11-07-92  16:29  440443e1   gc.txt
 ------          ------  ---                              -------
 157044           60266  62%                              2

$

script done on Sat Dec 12 00:39:50 1992

As you can see, there are no errors.

Keith
--
Keith Petersen
Maintainer of the MS-DOS archive at WSMR-SIMTEL20.Army.Mil [192.88.110.20]
Internet: [e-mail]     or      [e-mail]
Uucp: uunet!umich!vela!w8sdz                         BITNET: w8sdz@OAKLAND

------------------------------

Date: Fri, 11 Dec 1992 14:05:55 -0500 (EST)
From: Ivan Genekey Lau <[e-mail]>
Subject: PS/2 High Density mode for 3.6" Floppy isn't Compatible

It's simply because brand name IBM's use a disk drive that doesn't
search for the extra sense hole present in HD disks.  Thus, you can
have a low or high density disk and always be able to format it as HD.

On the other hand, clone machines almost always search for the extra
sense hold and won't read a disk as being high density unless it has
that hole.  A 3 1/2 disk formatted on an IBM as high density will not
be recognized on the clone machine unless the sensor is diabled.

------------------------------

Date: Fri, 11 Dec 92 13:51:30 EST
From: Ben John Fornshell <[e-mail]>
Subject: Winmarks, Specmarcs, and all that jazz

     In all the ratings that I've seen of PCs and various clones
thereof I have never seen a detailed description of what the various
ratings mean and was curious if anyone knew how they are derived, or
where I could find out that information.  I'm also curious as to
whether or not there is a comprehensive list of relative PC platforms,
for instance a comparison of everything from the 80086 to the 80586?

                                                Thanks in advance,
                                                [e-mail]

------------------------------

End of Info-IBMPC Digest V92 #196
*********************************
-------